Received: 26 September 2018 | Revised: 1 March 2019 | Accepted: 7 March 2019 DOI: 10.1111/eva.12793

ORIGINAL ARTICLE

A transcriptional and functional analysis of heat hardening in two invasive fruit fly species, Bactrocera dorsalis and Bactrocera correcta

Xinyue Gu1 | Yan Zhao1 | Yun Su1 | Jiajiao Wu2 | Ziya Wang1 | Juntao Hu3,4 | Lijun Liu1 | Zihua Zhao1 | Ary A. Hoffmann5 | Bing Chen6 | Zhihong Li1

1Department of Entomology, College of Plant Protection, China Agricultural Abstract University, Beijing, China Many insects have the capacity to increase their resistance to high temperatures by 2 Guangdong Inspection and Quarantine undergoing heat hardening at nonlethal temperatures. Although this response is well Technology Center, Guangzhou, China 3Redpath Museum, McGill University, established, its molecular underpinnings have only been investigated in a few species Montreal, Quebec, Canada where it seems to relate at least partly to the expression of heat shock protein (Hsp) 4 Department of Biology, McGill University, genes. Here, we studied the mechanism of hardening and associated transcription Montreal, Quebec, Canada responses in larvae of two invasive fruit fly species in China, Bactrocera dorsalis and 5School of BioSciences, Bio21 Institute, University of Melbourne, Parkville, Bactrocera correcta. Both species showed hardening which increased resistance to Victoria, Australia 45°C, although the more widespread B. dorsalis hardened better at higher tempera‐ 6State Key Laboratory of Integrated Management of Pest Insects and Rodents, tures compared to B. correcta which hardened better at lower temperatures. Institute of Zoology, Chinese Academy of Transcriptional analyses highlighted expression changes in a number of genes repre‐ Sciences, Beijing, China senting different biochemical pathways, but these changes and pathways were in‐ Correspondence consistent between the two species. Overall B. dorsalis showed expression changes Zhihong Li, Department of Entomology, College of Plant Protection, China in more genes than B. correcta. Hsp genes tended to be upregulated at a hardening Agricultural University, Beijing, China. temperature of 38°C in both species, while at 35°C many Hsp genes tended to be Email: [email protected] and upregulated in B. correcta but not B. dorsalis. One candidate gene (the small heat Bing Chen, State Key Laboratory of shock protein gene, Hsp23) with a particularly high level of upregulation was investi‐ Integrated Management of Pest Insects and Rodents, Institute of Zoology, Chinese gated functionally using RNA interference (RNAi). We found that RNAi may be more Academy of Sciences, Beijing, China. efficient in B. dorsalis, in which suppression of the expression of this gene removed Email: [email protected] and the hardening response, whereas in B. correcta RNAi did not decrease the hardening Ary A. Hoffmann, School of BioSciences, response. The different patterns of gene expression in these two species at the two Bio21 Institute, University of Melbourne, Parkville, Vic., Australia. hardening temperatures highlight the diverse mechanisms underlying hardening Email: [email protected] even in closely related species. These results may provide target genes for future

Present address control efforts against such pest species. Bing Chen, College of Life Sciences, Hebei University, Baoding, China KEYWORDS expression plasticity, hardening response, Hsp23, invasive species, thermal adaptation Funding information National Natural Science Foundation of China, Grant/Award Number: 31772230

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2019 The Authors. Evolutionary Applications published by John Wiley & Sons Ltd

 wileyonlinelibrary.com/journal/eva | 1147 Evolutionary Applications. 2019;12:1147–1163. 1148 | GU et al.

1 | INTRODUCTION offspring performance as well as sharing a common origin. However, the geographical distribution of B. dorsalis is now much wider than During the invasive and adaptive process, species often encounter that of B. correcta both in China and elsewhere in the invasive range, novel environmental conditions that require adaptation through which B. dorsalis has invaded including Hawaii, Kenya, and Tahiti and genetically based evolutionary changes, phenotypic plasticity or a gradually displaced pre‐established Ceratitis species in recent years combination of these processes (Gibert et al., 2016; Vázquez, Gianoli, (Figure 1; Hu, Chen, & Li, 2014; Liu et al., 2014; Liu, Jin, & Ye, 2013; Morris, & Bozinovic, 2017). Phenotypic plasticity or evolutionary Lux, Copeland, White, Manrakhan, & Billah, 2003; Reitz & Trumble, changes can help buffer organisms from environmental changes and 2002). Thus far, there are no records of B. dorsalis being displaced by thereby help their establishment and expansion, and even be the tar‐ other tephritid fly species (Liu et al., 2017). get of selection (Wellband & Heath, 2017). Many studies that con‐ Temperature tolerance may be one of the key factors influencing sider the ability of plastic changes to buffer environmental effects the distribution of Tephritidae like B. dorsalis and B. correcta (Hu et consider temperature extremes, which play an important role in the al., 2014; Liu & Ye, 2009; Pieterse, Terblanche, & Addison, 2017; Qin, success of the invasive process (David et al., 1997; Delpuech et al., Ni, et al., 2015; Qin, Paini, Wang, Fang, & Li, 2015). B. dorsalis and 1995; Klepsatel et al., 2013). B. correcta have similar cold tolerance to C. capitata that is known to Resistance to extreme high temperature represents a complex be capable of adapting to a wide range of climates; however, B. cor‐ of traits that have been strongly affected by the environment ex‐ recta is more susceptible to heat than B. dorsalis (Hallman, Myers, perienced previously (Wos & Willi, 2018). Heat hardening is one El‐Wakkad, Tadrous, & Jessup, 2013; Hu et al., 2014; Myers, Cancio‐ component of resistance, involving the rapid induction of protec‐ Martinez, Hallman, Fontenot, & Vreysen, 2016; Papadopoulos, tive biochemical and physiological mechanisms, which markedly 2008). This may reflect species differences in thermal tolerance, enhance heat resistance (Malmendal et al., 2006). This process mediated through transcription changes involving genes such as of heat hardening in insects is regarded to be related to poten‐ Hsp70 and Hsp90 (Hu et al., 2014). For instance in Liriomyza, Hsp tial molecules, physiological changes or the differential expres‐ gene expression at different temperatures in two Liriomyza likely in‐ sion of genes, such as the expression of heat shock protein (Hsp) fluenced their geographical distribution (Huang & Kang, 2007; King genes (Borchel, Komisarczuk, Rebl, Goldammer, & Nilsen, 2018; & MacRae, 2015). Dahlgaard, Loeschcke, Michalak, & Justesen, 1998; Manjunatha, However, mechanisms that underpin heat hardening in Bactrocera Rajesh, & Aparna, 2010; Sisodia & Singh, 2006; Sørensen, species and how they might contribute to differences among the Kristensen, & Loeschcke, 2003; Willot, Gueydan, & Aron, 2017). species are poorly characterized. As in other insects, there are lim‐ The induction of genes such as Hsp varies with the intensity of ited data on how species might differ in gene transcription under thermal hardening stress and the insect's physiological state (King hardening and how any differences relate to temperature adaptation & MacRae, 2015). In Drosophila, hardening following a nonlethal and organism performance (Clarke, 2003). In this paper, we consid‐ heat stress activates a heat shock response through altering the ered whether differences in temperature adaptability associated transcription and translation of a set of genes including small with hardening are linked to the regulation of certain genes or path‐ Hsp genes (sHsps) and other Hsps including Hsp70 (DiDomenico, ways, which in turn has contributed to the different distribution and Bugaisky, & Lindquist, 1982; Malmendal et al., 2006; Sørensen, Nielsen, Kruhøffer, Justesen, & Loeschcke, 2005). Other insects whose hardening responses have been characterized include the armyworm Mythimna separata, where preheating larvae enhances expression of genes encoding superoxide dismutase 1, catalase and Hsp70, the whitefly Bemisia tabaci, where Hsp23, 70 and 90 are upregulated, and the ant Cataglyphis mauritanica, where hard‐ ening is associated with the expression of two Hsc70‐4 cognates (Díaz, Orobio, Chavarriaga, & Toro‐Perea, 2015; Matsumura, Matsumoto, & Hayakawa, 2017; Willot et al., 2017). Bactrocera dorsalis (Hendel) and B. correcta (Bezzi) (Diptera: Tephritidae) are invasive pests that damage fruits and vegetables (Permpoon, Aketarawong, & Thanaphum, 2011). These species have been investigated because of their increasingly wide distributions and repeated invasions. The Bactrocera genus has a competitive advantage for oviposition over other fruit flies such as Ceratitis FIGURE 1 Global distribution map of Bactrocera correcta and B. dorsalis. The global distribution of the two species based on the capitata which is one of the most devastating and invasive world‐ information from GBIF (https://www.gbif.org/) and CABI (https:// wide pests (Liu, Zhang, Hou, Ou‐Yang, & Ma, 2017; Malacrida et www.cabi.org/). The blue dots and areas represent the distribution al., 2007). The two Bactrocera species are similar in many biologi‐ of B. correcta, and the red dots and areas represent the distribution cal attributes such as mating duration, oviposition preference, and of B. dorsalis GU et al. | 1149 invasive potential of the two species. We undertook a transcriptional chose 45°C for 1 hr as an extreme thermal stress which caused analysis to identify mechanisms underlying differences of heat hard‐ mortality in the range 40%–50% without heat hardening (Hu et ening responses in B. dorsalis and B. correcta. We first characterized al., 2014; Nyamukondiwa et al., 2011). After exposure to the heat the survival of these species under heat stress after a series of hard‐ stress, the larvae were returned to 25°C and their survival rate ening treatments. Then, we undertook a transcriptional analysis to was scored after 4 hr (Supporting Information Figure S1A). During identify key pathways and genes involved in hardening. Lastly, quan‐ the thermal treatments, samples were enclosed in 2 ml tubes with titative real‐time PCR (qRT‐PCR) and RNA interference (RNAi) were a hole in the lid, topped with 4 g diet to ensure the food supply used in a functional analysis of these two species around one gene and humidity. Tubes were suspended in a circulating water bath that seemed particularly important in the hardening response and set to the desired temperature. The control 25°C flies were also contributed to the different hardening response of these species. handled in the same way as the flies in the hardening experiments. Each replicate had 30 larvae in the 2 ml tube and each hardening treatment involved five biological replicates. 2 | MATERIALS AND METHODS

2.1 | Samples 2.3 | RNA extraction and cDNA synthesis

Bactrocera dorsalis and B. correcta were sourced from their first in‐ Individuals for mRNA sequencing, transcriptome verification and vaded range in China, from Guangdong province (N 23.40, E 113.22) the detection of RNA interference efficiency in each species were for B. dorsalis with the annual mean temperature 21.7°C and from collected randomly for RNA extraction. RNA was extracted from Yunnan province (N 23.60, E 102) for B. correcta with the annual whole body of third instar larvae using an RNA simple Total RNA kit mean temperature 25.8°C (Li, Wu, Chen, Wu, & Li, 2012; Liu & Ye, (Tiangen, China). cDNA was synthesized from 1,000 ng total RNA 2009). Cultures of these species were maintained in the laboratory using PrimeScript™ RT reagent kit with gDNA Eraser (Perfect Real for approximately 10 generations, across 3 years, at a temperature Time; Takara, Japan) following the manufacturer's instructions. of 25°C, a humidity of 70% and light period of 10D:14L in an en‐ vironment‐controlled incubator (Guo, Zhao, Liu, Li, & Shen, 2018; 2.4 | Transcriptome and transcriptome analysis of Nyamukondiwa, Terblanche, Marshall, & Sinclair, 2011; Weldon, heat hardening Nyamukondiwa, Karsten, Chown, & Terblanche, 2018). Cultures were maintained by turning over flies across three cages (each For mRNA sequencing, total RNA (5 μg) was isolated from three 45 cm × 45 cm × 50 cm), with 200 individuals per cage. To reduce groups of 30 3rd early‐instar larvae after exposing them to specific the risk of commonly observed inbreeding in fruit flies, we also regu‐ temperatures (25, 35 and 38°C) for 4 hr followed by 25°C for 1 hr larly initiated new cultures of the species from the same provinces for both B. dorsalis and B. correcta (Supporting Information Figure as above and added 30 new field fruit flies per cage from these cul‐ S1B) using an RNA simple Total RNA kit as described in Section 2.3. tures every half year (Hoffmann & Ross, 2018). Adult flies were given The quantity and integrity of the RNA were assessed with a Qubit® 25% sucrose and 75% peptone as their diet and immature stages RNA Assay kit in Qubit®2.0 Fluorometer (Life Technologies, CA, were cultured on an artificial diet described by Yuan et al. (2006). USA) and an RNA Nano 6000 Assay kit of the Agilent Bioanalyzer The most temperature‐sensitive development stage (7‐day‐old 3rd 2100 system (Agilent Technologies, CA, USA). For use as a template, early‐instar larvae) of these two species was chosen for hardening mRNA was enriched with a NEBNext Poly (A) mRNA Magnetic experiments (Hu et al., 2014; Jang, 1991). The 3‐day‐old 1st instar Isolation Module (NEB, E7490, Ipswich, MA, USA). RNA was chemi‐ larvae for hardening response study and 4‐day‐old 2nd instar larvae cally fragmented into 200–700 nt fragments and converted into for heat tolerance study of B. dorsalis and B. correcta were used in single‐stranded cDNA using random hexamer priming, followed by dsRNA‐feeding experiments. second‐strand cDNA synthesis using DNA Polymerase I and RNase H. The purified double‐stranded cDNA products were processed via magnetic beads, end repaired and ligated to adaptors. All li‐ 2.2 | Survival analysis under hardening braries were prepared using NEBNext®Ultra™ RNA Library Prep Five replicates with 30 larvae of each species, B. dorsalis and kit for Illumina® (NEB, USA) according to the manufacturer's pro‐ B. correcta, were exposed to heat hardening temperatures rang‐ tocol. Finally, after the assessment of fragment sizes through 2% ing from 34 to 40°C as described in Hu et al. (2014) for 4 hr in agarose gel electrophoresis and validation by quantitative real‐time a circulating water bath (PolyScience Programmable Temperature PCR using a library quantification kit/Illumina GA Universal (KAPA, Controller, Total Temperature Instrumentation, Inc., USA), while Wilmington, MA, USA), all libraries were sequenced on an Illumina the control group was maintained at 25°C. After pretreatment, HiSeqTM 2000 instrument (Illumina) at the Biomarker Technologies all tubes were returned to 25°C for 1 hr before exposure to heat company (Beijing, China). The clustering of index‐coded samples stress at 45°C for 1 hr in the water bath, which allowed vari‐ was performed on a cBot Cluster Generation System using TruSeq ous heat shock proteins and other genes to develop (Hoffmann, PE Cluster kit v3‐cBot‐HS (Illumina, San Diego, CA, USA), and Sørensen, & Loeschcke, 2003; Sørensen & Loeschcke, 2001). We paired‐end reads were generated through sequencing. 1150 | GU et al.

To remove adapter contamination, low‐quality bases and bases correlation coefficient (r) between the three biological replicates artificially introduced during library construction, we trimmed all (Schulze, Kanwar, Gölzenleuchter, Therneau, & Beutler, 2012). The raw reads using Trimmomatic 0.32 (http://www.usadellab.org/ analysis of differentially expressed genes (DEGs) was conducted cms/index.php?page=trimmomatic) with the following parame‐ in three steps. Firstly, read counts were adjusted by DESeq pack‐ ters: Phred33 LEADING: 3 TRAILING: 3 SLIDING WINDOW: 1:10 age through scaling normalized factor for each sequenced library MINLEN: 75 before transcript assembly, while the unpaired reads before analysis (Anders & Huber, 2010; Wang, Feng, Wang, Wang, were discarded (Bolger, Lohse, & Usadel, 2014). The clean reads & Zhang, 2009). Secondly, differential expression analysis of sam‐ were mapped to the sequences in the rRNA database of all published ples was performed using the DESeq (Anders & Huber, 2010; Wang insects downloaded from NCBI to discard rRNAs using SOAP (Li, Li, et al., 2009). In this analysis, p‐value adjusted by the Benjamini– Kristiansen, & Wang, 2008). Only the clean reads with the standard Hochberg approach was used to decrease false positives (Ferreira of Q30 > 85% and processed with Trimmomatic 0.32 and SOAP were & Zwinderman, 2006). Lastly, the standard fold change>|2| was set used for further analysis. As we wanted to compare the two species as the threshold for differential expression. To visualize transcrip‐ and only the genome for B. dorsalis was available (ASM78921v2; tion differences among treatments, a principal component analysis https://i5k.nal.usda.gov/Bactrocera_dorsalis), Trinity for research (PCA) was performed in R (2. 22. 0) with the prcomp function. This without the need for a genome sequence with set parameters (min_ was done with all the FPKM of DEG of three replicates in B. cor‐ kmer_cov: 2 min_contig_length: 200 group_pairs_distance: 500) was recta (Bc) and B. dorsalis (Bd) under three temperatures. Heat maps used for de novo transcriptome assembly to obtain corresponding of absolute DEG expression were generated as logarithmic values transcripts (Guo et al., 2018; Haas et al., 2013). The two species were that normalized FPKM for hardening temperatures (HT) by dividing DEG_FPKM of HT/DEG_FPKM of firstly assembled to get their own UniGene database, and then the by FPKM for control temperature (CK) (log10 general UniGene library was obtained by clustering the two individ‐ CK). GO enrichment analysis was implemented through the topGO R ual databases through CD‐Hit to compare the expression and ana‐ packages based a Kolmogorov–Smirnov test (Alexa & Rahnenfuhrer, lyze differences between the two species. Trinity was used to break 2010). KOBAS software was used to test statistically significant en‐ down clean reads into short fragments (Grabherr et al., 2011). Reads richment of differentially expressed genes in KEGG pathways (Xie of certain lengths of overlap were combined in contigs to form lon‐ et al., 2011). ger fragments. These assembled unigenes were further processed for sequence splicing and redundancy removal using sequence clus‐ 2.5 | Quantitative Real‐time PCR for tering software to acquire maximum length nonredundant unigenes transcriptome verification and the detection of RNA (Fu, Niu, Zhu, Wu, & Li, 2012). Briefly, Trinity was used for assembly interference efficiency to obtain corresponding transcripts, and we retained the transcript with the highest read coverage and removed the transcript with the In order to validate the results from the transcriptome analysis, lowest read coverage for each subcomponent. The transcripts se‐ we quantified the expression profiles of candidate genes Hsp23, lected in the clustering united as unigenes using the De Bruijn graph Hsp70 and Hsp90 for B. dorsalis and B. correcta. Thirty individuals algorithm and CD‐HIT to reduce sequence redundancy and improve for transcriptome verification (two species with three tempera‐ the performance of other sequence analyses (Fu et al., 2012; Yang ture treatments: 25, 35 and 38°C for 4 hr followed by 25°C for & Smith, 2013). 1 hr as described in Supporting Information Figure S1B) were col‐ For functional annotation, sequence alignments of unigenes to lected randomly. For the detection of RNA interference efficiency, the protein databases NR (nonredundant RefSeq proteins—NCBI), 10 individuals fed on dsRNA after 96 hr in each species were ran‐ Swiss‐Prot, GO (gene ontology), COG (Clusters of Orthologous domly selected for Hsp23 expression detection. RNA extraction Groups), KOG (euKaryotic Orthologous Groups), eggNOG4.5 and and cDNA synthesis followed the description in Section 2.3. The KEGG (Kyoto Encyclopedia of Genes and Genomes) were undertaken RNAi efficiency was quantified by quantitative real‐time PCR by using BLAST (E‐value < 1e‐5) (Altschul et al., 1997; Ashburner et using SYBR® Premix Ex Taq™ II (TliRNaseH Plus; Takara, Japan) al., 2000; Bairoch et al., 2005; Deng et al., 2006; Huerta‐Cepas et al., on an ABI 7500 instrument (Applied Biosystems Europe, Belgium). 2015; Kanehisa, Goto, Kawashima, Okuno, & Hattori, 2004; Tatusov, All RNA samples were performed with same methods and ana‐ Galperin, Natale, & Koonin, 2000). After predicting the amino acid lyzed in triplicate. The reaction included 1 μl cDNA, 12.5 μl SYBR sequence of unigenes through comparisons with the Pfam data‐ Green mix, 1 μl each of forward and reverse primers (10 p.m.), base, HMMER software was used to obtain annotation of protein 0.5 μl ROX Reference Dye II and 9 μl ddH2O. The thermocycler function (E‐value < 1e‐10) (Dahlgaard et al., 1998; Finn et al., 2013). conditions were 95°C for 30 s, followed by 40 cycles at 95°C for Bowtie was used to align the reads to the Trinity transcripts and es‐ 5 s and 60°C for 34 s. Melting curve analysis was performed at timate the number of RNA‐Seq fragments (Langmead, Trapnell, Pop, the end of each expression analysis, using the following condi‐ & Salzberg, 2009). With Bowtie alignment, FPKM (fragments per tions: 95°C for 15 s, followed by 60°C for 60 s with the decreas‐ kilobase of exon per million reads mapped) were quantified using ing rate of 1°C/s from 95°C. All the primers used are described in RSEM software (Li & Dewey, 2011). To evaluate the correlation of Supporting Information Table S4, and the amplification efficiency gene expression across different replicates, we calculated Pearson's of all the primers in these two species is close to 100% (Supporting GU et al. | 1151

Information Table S5; Hu et al., 2014; Shen, Huang, Jiang, Dou, using the T7 RiboMAX Express RNAi system (Promega). For larval & Wang, 2013). We evaluated the performance of five reference feeding, three grams artificial diet with 30 μl of a dsRNA solution genes, including 18s ribosomal RNA (18s), ribosomal protein L13 (1,000 ng/μl) was put into a 50 ml tube. Forty 3‐day‐old 1st instar (RPL13), succinate dehydrogenase (SD), α‐tubulin (α‐TUB) and β‐tu‐ larvae for hardening response study or 4‐day‐old 2nd instar larvae bulin (β‐TUB) in B. dorsalis and B. correcta using three software‐ for heat tolerance study of B. dorsalis and B. correcta were collected based approaches (BestKeeper, NormFinder and geNorm) and 18s and moved into a 50 ml tube with artificial diet, which had three was chosen as the reference gene in the experiment (Supporting holes in its lid for air. Five replicates per treatment were carried out Information Table S6). The qRT‐PCR data were analyzed using the and each replicate contained 40 larvae. For the temperature study, 2−ΔCT method (Chen & Wagner, 2012). Five biological replicates larvae were fed with dsHsp23 and dsGFP for 48 hr and transferred to were carried out for statistical analysis. All results from experi‐ new artificial diet for another 48 hr. After 96 hr feeding, the larvae mental replicates were analyzed with Student's t tests or one‐way developed to 3rd early‐ or late‐instar larvae. analyses of variance (ANOVAs) with SPSS 20 (IBM Corporation, USA). 2.9 | Extreme heat stress and hardening response after feeding dsHsp23 2.6 | Cloning the open reading frame of Hsp23 After 96 hr dsRNA feeding, 10 larvae of each species of B. dor‐ Hsp23 was selected as a candidate hardening gene based on the salis and B. correcta were killed for detecting Hsp23 gene expres‐ analysis of the transcriptomic data of B. dorsalis and B. correcta. sion. The rest of the 3rd late‐instar larvae were transferred to To verify the open reading frame (ORF) of Hsp23 in each spe‐ one 2 ml tube with diet for high temperature exposure. We chose cies, due to the high similarity on sequence, one pair of primers 3rd late‐instar larvae because of their high survival rate under ex‐ named Hsp23‐dsRNA‐F/R (Supporting Information Table S4) was treme temperature to study the function of Hsp23 on heat toler‐ designed using Primer Premier 5.0 software (PREMIER Biosoft ance (Jang, 1991). The larvae were directly exposed to heat stress International, USA) based on the sequence information from tran‐ at 45°C for 1 hr followed by 4 hr at room temperature before scriptome. The ORF and conserved domain were identified with scoring. As there was no response in the Hsp23‐knocking down the ORF Finder software (http://www.ncbi.nlm.nih.gov/gorf/gorf. in B. correcta at extreme high temperature, hardening response html). was only studied in B. dorsalis. Five groups of 40 3rd early‐instar larvae with low survival rate under extreme temperature were used to study the function of Hsp23 on heat hardening, which 2.7 | Phylogenetic analysis showed the recovery of Hsp23 expression could increase survival The integrity of homologous amino acid sequences of other species rate. After the feeding, the larvae were exposed to heat harden‐ was retrieved from the NCBI server (https://www.ncbi.nlm.nih.gov/; ing (35 and 38°C) and control (25°C) temperatures separately for Dou et al., 2017). Sequences were first aligned by the conserved se‐ 4 hr and returned to 25°C for 1 hr (Supporting Information Figure quences, and then, phylogenetic analysis was performed using the S1B). Ten of the flies were frozen in liquid nitrogen for the detec‐ neighbor‐joining method in the Molecular Evolutionary Genetics tion of Hsp23 expression. The other 30 flies were exposed to heat Analysis software (MEGA version 5.1; Tamura et al., 2011). All the stress at 45°C for 1 hr in a water bath (Supporting Information positions that contained gaps and missing data were eliminated be‐ Figure S1A). After heat stress, the larvae were moved back to fore alignment and phylogenetic analysis. 25°C and their survival rate scored after 4 hr. Each treatment was replicated five times.

2.8 | Validation of Hsp23 role through dsRNA 3 | RESULTS Hsp23 was chosen for functional analyses because it is the most expressed gene in B. dorsalis, the species with the stronger inva‐ 3.1 | Hardening response of B. dorsalis and sion potential, which likely involved the ability of this species to B. correcta adapt to increased temperature. While Hsp70 was the most ex‐ pressed gene in B. correcta, this gene was not selected for fur‐ We calculated the survival rate of 3rd instar larvae in the two species ther analysis given that it did not stand out in terms of expression (Figure 2). Both species responded to the hardening treatments from changes in B. dorsalis, whereas Hsp23 was upregulated in both spe‐ 34 to 40°C, with increased survival rate under heat stress of vary‐ cies (Figure 7). ing degrees. However, species had different hardening responses. Double‐stranded RNA of Hsp23 (dsHsp23) was used to knock Bactrocera dorsalis performed better than B. correcta in the 37–40°C down the expression of the target gene of B. dorsalis and B. correcta. hardening range, while B. correcta differed from B. dorsalis in the Green fluorescent protein (dsGFP) was used as a negative control. 34–36°C range. A hardening temperature of 35°C produced the larg‐ The dsRNA was synthesized by the primers named Hsp23‐dsR‐ est benefit for B. correcta, while 38°C led to the largest hardening NA‐F/R and Hsp23‐F/R‐T7 in Supporting Information Table S4 and response for B. dorsalis. 1152 | GU et al.

and B. correcta when compared to control treatments. In B. dorsalis, there were 430 upregulated genes and 303 downregulated genes (Figure 3b). There was relatively more upregulation in biological pro‐ cessing in cytochrome P450 of B. dorsalis, which reflected increas‐ ing P450 activities (Supporting Information Figures S3D and S4D). However, genes such as larval serum protein, dehydrogenase, ce‐ cropin and mucin were significantly downregulated. In B. correcta, 181 genes were upregulated and 201 genes were downregulated. Genes annotated with endocytosis, protein processing in endoplas‐ mic reticulum and spliceosome in KEGG analysis were upregulated (Supporting Information Figures S3A and S4A). Downregulated genes were annotated with peroxisome, glycolysis, pyruvate me‐ tabolism and biosynthesis of amino acids (Supporting Information FIGURE 2 Heat hardening response of Bactrocera correcta Figures S3A and S4A). These patterns indicated substantial changes and B. dorsalis. The survival rate of 7‐day‐old 3rd instar larvae in gene expression affecting a diversity of pathways. exposed to high temperature after heat hardening treatments. The temperatures used for hardening are indicated on the x‐axis. The In the comparison of 38 and 25°C, a total of 1,329 and 826 genes blue line represents the fitted curve of larval survival rate using showed significantly different expression in B. dorsalis and B. cor‐ an asymmetric model for B. correcta. The red line shows the fitted recta, respectively. In B. dorsalis, 815 genes were upregulated and curve for B. dorsalis using the same model. Different letters above 514 genes downregulated (Figure 3b). Genes related to lysosome, the bars indicate significant differences at p < 0.05, as determined glycolysis, carbon metabolism and biosynthesis of amino acids were by a Tukey HSD test. The capital letters in blue refer to B. correcta and the lowercase letters in red refer to B. dorsalis. The error bars upregulated at 38°C (Supporting Information Figures S3E and S4E). indicate 1 SE Genes related to carbon metabolism, metabolism of xenobiotics by cytochrome P450, protein processing in endoplasmic reticulum and pentose and glucuronate interconversions were significantly down‐ 3.2 | Transcriptome and bioinformatic analysis of regulated (Supporting Information Figures S3E and S4E). These B. dorsalis and B. correcta changes reflected energy metabolism such as glycolysis which may have provided energy for the hardening response, and protein pro‐ The percentage of bases with a Q30 score was over 92.64%, indicat‐ cessing in the endoplasmic reticulum which may affect the accumu‐ ing that the sequencing was reliable. We obtained 94,821 unigenes lation of unfolded proteins (Ron & Walter, 2007). In B. correcta, 435 from larval transcriptome sequencing, and a total of 42,931 unigenes genes were upregulated and 391 genes downregulated (Figure 3b). were annotated in the larvae transcriptome from databases in these Genes related to endocytosis, protein processing in endoplasmic two species (Supporting Information Tables S1 and S2). The length reticulum and spliceosome were upregulated at 38°C (Supporting distribution of the unigenes and statistics for larval unigenes assess‐ Information Figures S3B and S4B), which was similar to the response ment are shown in Supporting Information Figure S2 and Table S1, at 35°C. Genes such as peptidoglycan‐recognition protein, cuticle and the number of unigenes with a length >1,000 bp was 16,114. protein and lambda‐crystallin homolog were significantly downregu‐ The mapped ratio for all the samples was over 76% (Supporting lated. For the KEGG annotation analysis, the regulated genes mainly Information Table S3). In the transcriptome analysis, the PCA, bar involved the peroxisome, protein processing in endoplasmic reticu‐ charts and heatmaps of DGEs analyses revealed differences be‐ lum and metabolism of xenobiotics by cytochrome P450 (Supporting tween the Bactrocera species under different hardening tempera‐ Information Figures S3B and S4B), this suggests that heat stress may tures (Figures 3 and 4). In the PCA analysis of these two species, affect a range of oxidative processes. PC1 explained 77.3% of the variance in expression and clustered When comparing the two hardening temperature conditions B. dorsalis separately from B. correcta (Figure 3a). Although PC2, (35 vs. 38°C), 1,192 and 764 genes were detected with signifi‐ which explained 16.5% of the variance, clustered treatment samples cantly different expression in B. dorsalis and B. correcta, respectively of B. dorsalis separately, they showed no clear separation of B. cor‐ (Figure 3b). In contrast for B. dorsalis, a number of affected genes recta treatments (Figure 3a). More genes changed expression levels were involved in the biosynthesis of amino acids and carbon me‐ in B. dorsalis (Figure 3b) and expression patterns differed between tabolism (Supporting Information Figures S3F and S4F). There were the species as well as hardening temperatures (Figure 4). also many differentially expressed genes with unknown functions. In The variation in gene expression was analyzed in comparisons the KEGG analysis, most B. correcta genes related to the peroxisome, of 35 versus 25°C, 38 versus 25°C and 38 versus 35°C in the two protein processing in endoplasmic reticulum and fatty acid activities species. Compared to B. correcta, we found that there were more (Supporting Information Figures S3C and S4C). genes differentially expressed in B. dorsalis under both hardening There were more genes and processes affected at 38°C than at treatments (Figure 3b). For 35°C treatments, a total of 733 and 35°C (Figure 4a). Genes that showed certain expression patterns in 382 genes were differentially expressed, respectively, in B. dorsalis the two species formed 26 clusters (Figure 4b). Genes in cluster 2 GU et al. | 1153

FIGURE 3 Transcriptome analysis of Bactrocera correcta (Bc) and B. dorsalis (Bd) after exposure to different temperatures. (a) Clustering of control and heat hardening groups based on their transcriptome profiles by principal component analysis (PCA) across both species. Specific colors and points represent replicates for hardening treatments and species as shown in the legend. PCA was computed in R using the princomp function and a correlation matrix with the expression data of all the genes. Whereas B. correcta treatments formed a tight group, B. dorsalis was more variable depending on temperature. (b) Number of differentially expressed genes (DEGs) with fold change >2 for B. correcta and B. dorsalis. The chart above represents the number of DEGs at 35 versus 25°C, 38 versus 25°C and 38 versus 35°C in B. correcta, and the chart below represents the number of DEGs at 35 versus 25°C, 38 versus 25°C and 38 versus 35°C in B. dorsalis. Up‐ (red) and downregulated (green) unigenes are quantified and 9 annotated with immune response such as defense response to Hsp genes were upregulated with higher expression levels in B. cor‐ bacterium and innate immune response, stress response and some recta than in B. dorsalis, except for Hsp23, which had a very high ex‐ monooxygenase activities were upregulated only in B. correcta at pression level compared to all other genes under different hardening both hardening temperatures, while the related genes were down‐ conditions (Figure 5 and Supporting Information Figures S5 and S6 regulated in B. dorsalis. However, genes in cluster 13, 16, 17, 18, 20 and Table S7). Also, many other sHsps were associated with the hard‐ and 24 related to many developmental processes were only upreg‐ ening process, such as Hsp18.4 and Hsp20 (Supporting Information ulated in B. dorsalis, whereas genes in the same cluster 16, 17, 24 Figures S5 and S6 and Table S7). were downregulated in B. correcta. For the KEGG terms, enriched Although there were common genes upregulated by both pathways numbered 89 in the 35 versus 25°C comparison and 123 hardening treatments, Hsp23 was the only annotated Hsp gene in the 38 versus 25°C comparison for B. dorsalis, and the equivalent upregulated in all combinations of temperature groups in B. dor‐ numbers for B. correcta were 63 (35 vs. 25°C) and 93 (38 vs. 25°C; salis through Wayne maps (Figure 6c). The same Hsp23 gene, Figure 4c). among several Hsp genes, was also significantly upregulated in Differentially expressed genes including up‐ and downregulated B. correcta (Figure 6a). Among all the genes upregulated under the genes at different temperature treatments remained positively cor‐ two hardening temperatures (3 genes in B. dorsalis and 7 genes related in expression levels between treatments in the scatter plots in B. correcta), the Hsps were the only BLAST genes identified as (Figure 5). As Hsps were the largest family responding to stress clear candidates. Some Hsp genes were downregulated, including among all the DEGs in these two species with the highest levels of Hsp83, which was downregulated when compared to the control expression (Figure 5) and Hsp genes such as Hsp70, Hsp90 and small groups under two hardening temperatures in these two species heat shock protein (sHsp) have been reported as playing roles in high (Figure 6b,d and Supporting Information Table S7). In comparison temperature responses previously (Borchel et al., 2018; Dahlgaard et with the other genes including the Hsp genes, the expression of al., 1998; Manjunatha et al., 2010; Sisodia & Singh, 2006; Sørensen Hsp23 was at a high level at all temperatures (Figure 5) and as‐ et al., 2003; Willot et al., 2017), we considered Hsps separately. More sociated with the weight of early pupae in B. dorsalis (Supporting 1154 | GU et al.

FIGURE 4 Heatmaps of DEGs under different hardening treatments (35, 38 and 25°C control) in Bactrocera correcta (Bc) and B. dorsalis (Bd). (a) The heatmap of DEGs in B. correcta and B. dorsalis. Differential expression is shown as log10 fold changes (LFC) of hardening temperatures versus 25°C using the FPKM for hardening temperatures/ value of log10 FPKM for control temperature. For clarity, LFC values were capped at 1.5 and −1.5. The color bars besides the heatmaps indicate the relative size of the changes. (b) Differential expression patterns between the two species across treatments. (c) Classification of KEGG pathways under different hardening treatments. The number below the bars represents the number of enriched KEGG pathways for each experimental group

Information Figure S7). In general, Hsp23 is one of the highest ex‐ 3.5 | Transcriptome verification of Hsp23 in pressed genes in all DEGs including Hsps in both species (Figure 5). B. correcta and B. dorsalis

Compared with other two Hsps, Hsp70 and Hsp90, that are po‐ 3.3 | Cloning and characterization of Hsp23 tentially related to temperature adaptation in B. correcta and Hsp23 from both species was cloned from cDNA. Based on the phy‐ B. dorsalis, Hsp23 increased significantly at the two hardening logenetic analysis and sequencing results, BcHsp23 and BdHsp23 temperatures and had expression patterns across samples in these were similar to each other, with a difference in nucleic acid and two species consistent with the transcriptome results (Figure 7; amino acid sequences of ORFs of only 17 out of 513 base pairs and Hu et al., 2014). The expression level of Hsp23 was similar under 7 out of 170 amino acids. 35°C in two species but showed a difference at 38°C. Compared to the control group, the fold change in B. correcta was 2.4 and 4.6 times for 35 and 38°C, respectively (Figure 7a). This compared 3.4 | Phylogenetic analysis of sHsp from different to fold changes in B. dorsalis of 5.3 and 43.1 times, respectively insect species (Figure 7b). In these two species, the other two Hsps had the simi‐ The NJ tree (Supporting Information Figure S8) showed two annotated lar expression pattern; Hsp70 expression was increased at 38°C Hsp23 genes in B. correcta and B. dorsalis were clustered together, with but there was no change in Hsp90 at the two hardening tempera‐ high sequence similarity to the Hsp23 group from other species. tures (Figure 7). GU et al. | 1155

FIGURE 5 Relationship between mRNA levels for genes and heat shock protein genes (Hsps) with fold change more than 2 under hardening/control treatment comparisons (35, 38 and 25°C controls). DEGs are measured as a logarithmic value of mean FPKM for three mean FPKM replicates (log10 ) in Bactrocera correcta (Bc) and B. dorsalis (Bd). Transcripts encoding Hsps are shown as purple triangles, Hsp23 is highlighted in red (arrow), and all other transcripts are represented by gray circles. The black line represents k = 1, and the red line and zone (±95% confidence bands) represent the least‐squares regression fit of Hsp transcripts, for which Hsp23 was a significant outlier in B. dorsalis under the two hardening temperatures when compared to 25°C

group, the fold change expression of Hsp23 in B. dorsalis was 2.87 3.6 | Gene function and 23.61 times for 35 and 38°C, respectively (Figure 8c), and the Bactrocera dorsalis has been suggested to have a stronger invasion survival rate rose to 90.65% and 92.71% for 35 and 38°C, respec‐ ability compared with other tephritid fruit flies, and we chose Hsp23 tively (Figure 8d). The survival rate of the dsGFP group therefore in‐ in this species to study whether this gene influenced temperature ad‐ creased significantly under both hardening temperatures compared aptation and caused the difference between these two species (Liu to the control groups. But for the dsHsp23‐feeding group, compared et al., 2017; Malacrida et al., 2007). Despite the similar expression to the treatments under 25°C, survival rate only increased under pattern and sequence, the gene showed distinct functions between 38°C when expression level of the target gene also significantly these two species. Therefore, an exposure of 3rd late‐instar larvae increased. For B. correcta, although the target gene Hsp23 showed was used to study whether low expression of Hsp23 could decrease a 0.60‐fold decrease in expression (Figure 8e), survival rate under survival. Compared with the dsGFP‐feeding group, dsHsp23 expo‐ 45°C for 1 hr was not affected (Figure 8f). sure led to a 0.48‐fold reduction in expression in B. dorsalis (Figure 8a) while survival rate under 45°C was decreased by 42.9% (Figure 8b). 4 | DISCUSSION The result suggests that constitutive expression of Hsp23 increases survival of B. dorsalis under heat. We used two hardening tempera‐ 4.1 | Species differences in hardening tures of 35 and 38°C to study whether suppressed Hsp23 expression influenced survival rate following hardening. In the dsHsp23‐feeding We showed that while both Bactrocera species were hardened treatments, compared to the 25°C group, fold change expression of by nonlethal temperatures, the temperatures range of 37–40°C Hsp23 in B. dorsalis increased by 1.96 and 83.27 times for the 35 and led to a stronger hardening response for B. dorsalis, while B. cor‐ 38°C treatments, respectively (Figure 8c), while survival rate rose to recta performed better following exposure to 34–36°C (Figure 2). 47.05% and 82.36%, respectively (Figure 8d). For the dsGFP‐feeding The more narrowly distributed B. correcta seemed to show a heat 1156 | GU et al.

these two species (Hu et al., 2014; Liu & Ye, 2009; Pieterse et al., 2017). For B. dorsalis, heat shock tolerance of larvae was signifi‐ cantly enhanced by exposing to heat hardening 37 and 39°C for 1 or 2 hr (Hu et al., 2014; Pieterse et al., 2017). In B. correcta, the temperatures from 30 to 33°C appeared to be the most suitable for egg, larva and pupa development, and the preoviposition time was shortened following exposure to 33–36°C (Liu & Ye, 2009). Whether these differences in hardening responses contribute to differences in the distribution of the two species is unclear. Other comparisons of insects also provide mixed evidence for as‐ sociations between hardening and species distributions. In a com‐ parison of the widely distributed C. capitata with its more narrowly distributed relative C. rosa, both species show similar levels of survival under acute high and low temperatures exposures when reared under common conditions and with a pretreatment of 36°C (1 hr) which altered survival at 41°C (Nyamukondiwa, Kleynhans, & Terblanche, 2010). Rapid cold hardening ability has been linked to species distributions in Collembola, and heat hardening has been related to different distributions in two Cataglyphis ants (Bahrndorff, Loeschcke, Pertoldi, Beier, & Holmstrup, 2009; Willot et al., 2017), but there is often no link between hardening and spe‐ cies distributions in Drosophila (Overgaard, Kristensen, Mitchell, & Hoffmann, 2011). Apart from hardening responses, other environ‐ mental factors can also influence the distribution and abundance of tephritid species (Vayssières, Carel, Coubes, & Duyck, 2008). Some factors affect the fitness of fruit flies and have an indirect influence on their distribution, while others have a direct influence. It was found that high temperatures, low humidity and the absence of suitable host fruits for oviposition could cause ovarian imma‐ turity thus influence further spread and distribution of Bactrocera species (De Meyer et al., 2010). Due to the preference for certain hosts, Dacus ciliatus enhances its biotic potentialities and maintains its population at low levels especially at low altitudes to avoid com‐ petition with the melon fly B. cucurbitae (Vayssières et al., 2008). For Bactrocera species, compared to B. dorsalis, B. correcta sur‐ vival was more stable across all the high temperatures, which meant FIGURE 6 Wayne maps of up‐ (red, light green, yellow and this species was not as sensitive as B. dorsalis to the hardening purple) and downregulated (blue, dark green, light and dark temperatures. The transcriptome results, including the number of gray) DEGs and Hsp genes that were differentially expressed regulated genes and pathways, also showed the stability of hard‐ (35 vs. 25 , 38 vs. 25 , 38 vs. 35 ) in Bactrocera correcta and ℃ ℃ ℃ ening response. The stable environment in the distribution range of B. dorsalis. (a) The upregulated DEGs and Hsp genes in B. correcta. B. correcta such as Yunnan, China, might contribute to this situation. (b) The downregulated DEGs and Hsp genes in B. correcta. (c) The upregulated DEGs and Hsp genes in B. dorsalis. (d) The Stability of transcription might bring some advantages such as a sta‐ downregulated DEGs and Hsp genes in B. dorsalis. Bold numbers in ble metabolic rate when temperature ranges around mild‐high lev‐ parentheses represent total numbers of regulated unigenes in these els, especially from 34 to 36°C. This stability may facilitate the ability two species of B. correcta to invade into areas that lack extreme heat. Given its rapid hardening response reflected by survival rate, B. dorsalis may hardening response more rapidly than the more widely distrib‐ be able to invade wider areas, especially under the climate change. uted species. These results highlight differences among species in hardening responses which has also been noted in other groups of 4.2 | Gene expression patterns under two hardening related species, particularly in Drosophila (Malmendal et al., 2006; temperatures Nyamukondiwa et al., 2011; Overgaard, Sørensen, Com, & Colinet, 2014; Willot et al., 2017). They also appear to be consistent with Elevated temperature exposure elicited a significant change in gene previous studies on temperature responses of other life stages of expression profiles between B. dorsalis and B. correcta. Many of the GU et al. | 1157

FIGURE 7 Expression of Hsp23, Hsp70 and Hsp90 at two hardening temperatures (35 and 38°C) and the 25°C control in Bactrocera correcta (Bc) and B. dorsalis (Bd). (a) Expression of Hsp23, Hsp70 and Hsp90 in B. correcta. (b) Expression of Hsp23, Hsp70 and Hsp90 in B. dorsalis. 18s was used as the control gene. Different letters above the bars represent significant differences at p < 0.05, as determined by a t test differences we identified correspond to specific genes or functional 6a,c). However, in B. dorsalis, most Hsp genes were downregulated categories that have relevance to thermal tolerance in these two at 35°C except Hsp23. We checked heat shock factors and thermal species, including genes such as Hsps, antioxidants/oxidative stress receptors including transient receptor potential (TRP), ion trans‐ enzymes and pathways like lysosome, glycolysis, carbon metabo‐ porter, Gr28b.d and Na+/ K+‐ATPase in the list of differentially lism and biosynthesis of amino acids as mentioned in the Results. expressed genes. Only one Na+P‐type‐ATPase was found down‐ In the PCA analysis and the comparison of up‐ and downregulated regulated in the two Bactrocera species at 35°C, which may reflect genes (Figure 3), it appears that the transcriptome of B. dorsalis is a decrease of Na+ transportation and signaling transduction. Such more responsive to high temperature than B. correcta, especially differences may contribute to different hardening responses of under 38°C, which contributed to a more variable survival rate. the species at the two temperatures. In other studies using re‐ Interestingly, B. correcta shows a more rapid hardening response yet lated species, differences in expression level of certain genes in its transcriptome is not as responsive and the genes in B. correcta response to the same environmental conditions have also been regulated more stably, which was consistent with survival rate. The noted, including in fish, corals and planthoppers (Barshis et al., different responses between species were also evident in our GO 2013; Huang et al., 2017; Wellband & Heath, 2017). and KEGG analysis (Supporting Information Figures S3 and S4), with the regulated genes in B. correcta annotated with protein synthesis 4.3 | Hsp23 expression and heat tolerance that could protect organisms against heat and oxidative stresses by eliminating hydrogen peroxide (Lu, Bai, Zheng, & Lu, 2017), whereas We showed that suppressing Hsp23 expression decreased the in B. dorsalis, changes to genes involved in a range of metabolic pro‐ hardening response of B. dorsalis. However, RNAi induced differ‐ cesses such as carbon metabolism were involved under hardening ent responses between B. dorsalis and B. correcta. In many studies, temperatures (Figure 4). Hsp70 has been suggested essential to heat tolerance (Bahrndorff, When considering both species, the response to stimulus was Mariën, Loeschcke, & Ellers, 2009; Bettencourt, Hogan, Nimali, one of the most significantly enriched GO terms considering the & Drohan, 2008; Dahlgaard et al., 1998; Sørensen & Loeschcke, DEG number and ratio of DEGs of the inquiry set in all related 2001; Zizzari & Ellers, 2011). However, in our study, we found that genes of the reference set (Supporting Information Figure S3). expression of Hsp23 changed more rapidly and more intensely Hsps were the largest family to respond to stimulus among all the than Hsp70 in B. dorsalis, which might be caused by sHsps at low DEGs, with the highest levels of expression (Figure 5). Among all Hsp70 concentrations inhibiting the disaggregation and refolding the genes upregulated in two hardening temperatures, the Hsps processes (Żwirowski et al., 2017). As sHsps act as the first line of were the only ones which were clearly candidates for the harden‐ defense and are the key factors in modifying protein aggregation, ing process. Despite differences among species in some groups of the rapid expression of Hsp23 indicates that the sHsp can respond genes, both species showed a number of upregulated Hsp genes to temperature change more quickly than other Hsps (Żwirowski associated with the hardening response under 38°C (Figures 5 and et al., 2017). 1158 | GU et al.

FIGURE 8 Functional analysis of Hsp23 using RNAi in Bactrocera dorsalis and B. correcta. (a) Expression of Hsp23 after RNAi in B. dorsalis. (b) Survival rate of B. dorsalis under the extreme thermal stress of 45°C for 1 hr. (c) Expression of Hsp23 after heat hardening temperatures in B. dorsalis. (d) Survival rate of dsRNA‐feeding larvae exposed to extreme thermal stress 45°C after heat hardening treatments in B. dorsalis. (e) Expression of Hsp23 after RNAi in B. correcta. (f) Survival rate of B. correcta under extreme thermal stress 45°C for 1 hr. All the flies of B. correcta and B. dorsalis were fed by dsHsp23 and dsGFP for 96 hr. 18s was used as the control gene. The letter “*” and “**” above the bars represent significant differences at p < 0.05 and p < 0.01, respectively, and “ns” represents no significant differences. Different letters above indicate significant differences at p < 0.05, as determined by a t or Tukey HSD test

While Hsp23 which was not the most expressed gene in B. cor‐ pattern and sequences (Figure 7). The sHsps play important roles in recta, it was chosen for functional studies because of the strong inva‐ many key physiological activities such as protein folding and trans‐ sion ability in B. dorsalis where it was the most expressed gene, and portation, embryo development, and immunization mechanisms (Li considered in both species because of high similarities in expression et al., 2009). In a recent study, binding by these proteins has been GU et al. | 1159 suggested being important mechanisms for protecting other cellular biosynthesis. As more genes and pathways are regulated in B. dorsalis, proteins from denaturation under thermal stress and other stresses, this species may be more adaptable under high temperature stress. and overexpression of sHsps can enhance the tolerance of cells to temperature changes (Li et al., 2009; Wang et al., 2017). Among sHsps, Hsp23 shows a key role in temperature adaptation in many 5 | CONCLUSIONS species. In Drosophila melanogaster, gene knockdown experiments have suggested that Hsp22 and Hsp23 genes contribute to adaptive It is important to understand the role of genes such as Hsps and responses to fluctuating thermal conditions and particularly in chill pathways in relation to the hardening response and stress resist‐ coma recovery (Colinet, Lee, & Hoffmann, 2010). Also, overexpres‐ ance, in order to develop a mechanistic basis to evolutionary change sion of Hsp23 muscle‐specific has been suggested to promote pro‐ and plastic responses (Sørensen et al., 2003). In our study, we used teostasis and protect muscle from heat stress (Kawasaki et al., 2016). two successful invasive species that have invaded different thermal Moreover, the overexpression of muscle‐specific Hsp23 gene in fe‐ environments to understand the role of hardening and plasticity male ovaries produced offspring embryos with increased thermal in climate adaptation. Based on our results, the Bactrocera species tolerance (Lockwood, Julick, & Montooth, 2017). In other species have a varied ability to adapt to temperatures. Different harden‐ such as Leishmania donovani, the scuttle fly Megaselia scalaris and ing responses may be related to the regulation of certain genes like the whitefly Bemisia tabaci, Hsp23 expression is required for surviv‐ Hsp23 which in turn could contribute to species distributions and ing extreme temperature treatments (Díaz et al., 2015; Hombach, invasive potential. The widely distributed species, B. dorsalis, seems Ommen, MacDonald, & Clos, 2014; Malewski et al., 2015). more sensitive to temperature change at the transcriptome level. We used hardening temperatures 35 and 38°C to increase This sensitivity may help adaptive thermal responses, whereas the expression of Hsp23. Although we showed an effect of Hsp23 B. correcta might be at an advantage in milder environments. expression knockdown on the heat tolerance and hardening of However, this study only provides a starting point for understand‐ B. dorsalis, we did not find a similar effect in B. correct on high ing the genomic basis of climate adaptation in invasive fruit flies and temperature tolerance. Perhaps a higher level of knockdown is the functional studies in particular need to be expanded to other required to trigger an effect of this gene on tolerance in B. cor‐ gene families. Eventually, these types of studies may indicate tar‐ recta, particularly if the gene has multiple functions and there is a gets for genetic manipulation to eventually control these pests. For cost of gene expression associated with hardening on other fitness instance, the importance of Hsp23 in heat hardening in B. dorsalis components (Chen, Feder, & Kang, 2018; Sørensen et al., 2003). may point to this gene as being a useful candidate for gene drive It is also possible that this gene has no impact on the hardening technology to suppress this pest at warm times of the year. response of B. correcta, despite the substantial upregulation of this gene under the hardening treatment. We did not consider the as‐ ACKNOWLEDGEMENTS sociation between this gene and the hardening response in B. cor‐ recta for three reasons. Firstly, despite a high level of sequence The authors would like to thank Mr. Guoping Zhan and Mr. Chen conservation between the species, the thermal responses did not Ma in Chinese Academy of Inspection and Quarantine for provid‐ decrease at 45°C even with successful knockdown as we men‐ ing samples of B. dorsalis and B. correcta. This study was financially tioned in the methods, suggesting that Hsp23 has limited or no ef‐ supported by the National Natural Science Foundation of China fect on heat tolerance (Figure 8). Secondly, according to the similar (no. 31772230). expression level of Hsp23 induced at 35°C in B. dorsalis and B. cor‐ recta, two hardening temperatures (35 and 38°C) after the RNAi in CONFLICT OF INTEREST B. correcta might not recover the expression of this gene (Figures 7 and 8). Finally, although Hsp23 showed an increased expression None declared. level, there was no significant difference in survival rate between

35 and 38°C (Figures 2 and 7a). DATA ACCESSIBILITY Ambient temperature is a key environmental factor influencing the ecology and evolution of ectotherms, and invasive species can be RNA‐Seq data are available at National Center for Biotechnology used as models for investigating issues related to adaptive processes Information (NCBI): Sequence Read Archives SRP158095 and under changing ambient conditions and bringing ecological and evo‐ BioProject accession PRJNA486250. The data sets supporting lutionary processes into a common framework (Bennett & Lenski, the conclusions of this article are included within the Supporting 2007; Diamond, Chick, Perez, Strickler, & Martin, 2017; Ghalambor, Information files. McKay, Carroll, & Reznick, 2007). In our research, despite being similar, these two invasive species differ in many respects such as ORCID heat hardening profiles, transcriptome responses and Hsp expres‐ sion. These differences likely mediate thermal hardening and involve Xinyue Gu https://orcid.org/0000-0003-0591-0356 genes such as Hsps and pathways such as energy metabolism and Bing Chen https://orcid.org/0000-0002-1238-3948 1160 | GU et al.

REFERENCES David, J. R., Gibert, P., Gravot, E., Petavy, G., Morin, J. P., Karan, D., & Moreteau, B. (1997). Phenotypic plasticity and developmental tem‐ Alexa, A., & Rahnenfuhrer, J. (2010). topGO: Enrichment analysis for gene perature in Drosophila: Analysis and significance of reaction norms ontology. R package version 2.22. Vienna, Austria: R Foundation for of morphometrical traits. Journal of Thermal Biology, 22(6), 441–451. Statistical Computing Vienna. https://doi.org/10.1016/S0306-4565(97)00063-6 Altschul, S. F., Madden, T. L., Schäffer, A. A., Zhang, J., Zhang, Z., Miller, De Meyer, M., Robertson, M. P., Mansell, M. W., Ekesi, S., Tsuruta, K., W., & Lipman, D. J. (1997). Gapped BLAST and PSI‐BLAST: A new gen‐ Mwaiko, W., … Peterson, A. T. (2010). Ecological niche and potential eration of protein database search programs. Nucleic Acids Research, geographic distribution of the invasive fruit fly Bactrocera invadens 25(17), 3389–3402. https://doi.org/10.1093/nar/25.17.3389 (Diptera, Tephritidae). Bulletin of Entomological Research, 100(1), 35– Anders, S., & Huber, W. (2010). Differential expression analysis for 48. https://doi.org/10.1017/S0007485309006713 sequence count data. Genome Biology, 11(10), R106. https://doi. Delpuech, J. M., Moreteau, B., Chiche, J., Pla, E., Vouidibio, J., & David, org/10.1186/gb-2010-11-10-r106 J. R. (1995). Phenotypic plasticity and reaction norms in temperate Ashburner, M., Ball, C. A., Blake, J. A., Botstein, D., Butler, H., Cherry, J. and tropical populations of Drosophila melanogaster: Ovarian size and M., … Sherlock, G. (2000). Gene ontology: Tool for the unification developmental temperature. Evolution, 49(4), 670–675. of biology. Nature Genetics, 25(1), 25. https://doi.org/10.1038/75556 Deng, Y., Li, J., Wu, S., Zhu, Y., Chen, Y., & He, F. (2006). Integrated nr Bahrndorff, S., Loeschcke, V., Pertoldi, C., Beier, C., & Holmstrup, M. database in protein annotation system and its localization. Computer (2009). The rapid cold hardening response of Collembola is influ‐ Engineering, 32(5), 71–74. enced by thermal variability of the habitat. Functional Ecology, 23(2), Diamond, S. E., Chick, L., Perez, A., Strickler, S. A., & Martin, R. A. (2017). 340–347. Rapid evolution of ant thermal tolerance across an urban‐rural tem‐ Bahrndorff, S., Mariën, J., Loeschcke, V., & Ellers, J. (2009). Dynamics of perature cline. Biological Journal of the Linnean Society, 121(2), 248– heat‐induced thermal stress resistance and hsp70 expression in the 257. https://doi.org/10.1093/biolinnean/blw047 springtail, Orchesella cincta. Functional Ecology, 23(2), 233–239. Díaz, F., Orobio, R. F., Chavarriaga, P., & Toro‐Perea, N. (2015). Differential Bairoch, A., Apweiler, R., Wu, C. H., Barker, W. C., Boeckmann, B., Ferro, expression patterns among heat‐shock protein genes and thermal re‐ S., … Martin, M. J. (2005). The universal protein resource (UniProt). sponses in the whitefly Bemisia tabaci (MEAM 1). Journal of Thermal Nucleic Acids Research, 33(suppl_1), D154–D159. Biology, 52, 199–207. https://doi.org/10.1016/j.jtherbio.2015.07.004 Barshis, D. J., Ladner, J. T., Oliver, T. A., Seneca, F. O., Traylor‐Knowles, DiDomenico, B. J., Bugaisky, G. E., & Lindquist, S. (1982). Heat shock N., & Palumbi, S. R. (2013). Genomic basis for coral resilience to cli‐ and recovery are mediated by different translational mechanisms. mate change. Proceedings of the National Academy of Sciences, 110(4), Proceedings of the National Academy of Sciences, 79(20), 6181–6185. 1387–1392. https://doi.org/10.1073/pnas.1210224110 https://doi.org/10.1073/pnas.79.20.6181 Bennett, A. F., & Lenski, R. E. (2007). An experimental test of evolution‐ Dou, W., Tian, Y., Liu, H., Shi, Y., Smagghe, G., & Wang, J. J. (2017). ary trade‐offs during temperature adaptation. Proceedings of the Characteristics of six small heat shock protein genes from Bactrocera National Academy of Sciences, 104(suppl 1), 8649–8654. https://doi. dorsalis: Diverse expression under conditions of thermal stress org/10.1073/pnas.0702117104 and normal growth. Comparative Biochemistry and Physiology Bettencourt, B. R., Hogan, C. C., Nimali, M., & Drohan, B. W. (2008). Part B: Biochemistry and Molecular Biology, 213, 8–16. https://doi. Inducible and constitutive heat shock gene expression responds to org/10.1016/j.cbpb.2017.07.005 modification of Hsp70 copy number in Drosophila melanogaster but Ferreira, J., & Zwinderman, A. (2006). On the Benjamini‐Hochberg does not compensate for loss of thermotolerance in Hsp70 null flies. method. Annals of Statistics, 34(4), 1827–1849. https://doi. BMC Biology, 6(1), 5. https://doi.org/10.1186/1741-7007-6-5 org/10.1214/009053606000000425 Bolger, A. M., Lohse, M., & Usadel, B. (2014). Trimmomatic: A flexible Finn, R. D., Bateman, A., Clements, J., Coggill, P., Eberhardt, R. Y., Eddy, trimmer for Illumina sequence data. Bioinformatics, 30(15), 2114– S. R., … Mistry, J. (2013). Pfam: The protein families database. Nucleic 2120. https://doi.org/10.1093/bioinformatics/btu170 Acids Research, 42(D1), D222–D230. Borchel, A., Komisarczuk, A. Z., Rebl, A., Goldammer, T., & Nilsen, F. Fu, L., Niu, B., Zhu, Z., Wu, S., & Li, W. (2012). CD‐HIT: Accelerated (2018). Systematic identification and characterization of stress‐in‐ for clustering the next‐generation sequencing data. Bioinformatics, ducible heat shock proteins (HSPs) in the salmon louse (Lepeophtheirus 28(23), 3150–3152. https://doi.org/10.1093/bioinformatics/bts565 salmonis). Cell Stress and Chaperones, 23(1), 127–139. https://doi. Ghalambor, C. K., McKay, J. K., Carroll, S. P., & Reznick, D. N. (2007). org/10.1007/s12192-017-0830-9 Adaptive versus non‐adaptive phenotypic plasticity and the Chen, B., Feder, M. E., & Kang, L. (2018). Evolution of heat‐shock protein potential for contemporary adaptation in new environments. expression underlying adaptive responses to environmental stress. Functional Ecology, 21(3), 394–407. https://doi.org/10.1111/j.1365- Molecular Ecology, 27(15), 3040–3054. https://doi.org/10.1111/ 2435.2007.01283.x mec.14769 Gibert, P., Hill, M., Pascual, M., Plantamp, C., Terblanche, J. S., Yassin, A., Chen, B., & Wagner, A. (2012). Hsp90 is important for fecundity, lon‐ & Sgrò, C. M. (2016). Drosophila as models to understand the adap‐ gevity, and buffering of cryptic deleterious variation in wild fly tive process during invasion. Biological Invasions, 18(4), 1089–1103. populations. BMC Evolutionary Biology, 12(1), 25. https://doi. https://doi.org/10.1007/s10530-016-1087-4 org/10.1186/1471-2148-12-25 Grabherr, M. G., Haas, B. J., Yassour, M., Levin, J. Z., Thompson, D. A., Clarke, A. (2003). Costs and consequences of evolutionary temperature Amit, I., … Regev, A. (2011). Full‐length transcriptome assembly from adaptation. Trends in Ecology & Evolution, 18(11), 573–581. https:// RNA‐Seq data without a reference genome. Nature Biotechnology, doi.org/10.1016/j.tree.2003.08.007 29(7), 644–652. https://doi.org/10.1038/nbt.1883 Colinet, H., Lee, S. F., & Hoffmann, A. (2010). Functional characterization Guo, S., Zhao, Z., Liu, L., Li, Z., & Shen, J. (2018). Comparative transcrip‐ of the Frost gene in Drosophila melanogaster: Importance for recovery tome analyses uncover key candidate genes mediating flight capac‐ from chill coma. PLoS ONE, 5(6), e10925. https://doi.org/10.1371/ ity in Bactrocera dorsalis (Hendel) and Bactrocera correcta (Bezzi) journal.pone.0010925 (Diptera: Tephritidae). International Journal of Molecular Sciences, Dahlgaard, J., Loeschcke, V., Michalak, P., & Justesen, J. (1998). Induced 19(2), 396. thermotolerance and associated expression of the heat‐shock pro‐ Haas, B. J., Papanicolaou, A., Yassour, M., Grabherr, M., Blood, P. D., tein Hsp70 in adult Drosophila melanogaster. Functional Ecology, 12(5), Bowden, J., … Regev, A. (2013). De novo transcript sequence recon‐ 786–793. https://doi.org/10.1046/j.1365-2435.1998.00246.x struction from RNA‐seq using the Trinity platform for reference GU et al. | 1161

generation and analysis. Nature Protocols, 8(8), 1494. https://doi. Li, R., Li, Y., Kristiansen, K., & Wang, J. (2008). SOAP: Short oligonucle‐ org/10.1038/nprot.2013.084 otide alignment program. Bioinformatics, 24(5), 713–714. https://doi. Hallman, G. J., Myers, S. W., El‐Wakkad, M. F., Tadrous, M. D., & Jessup, org/10.1093/bioinformatics/btn025 A. J. (2013). Development of phytosanitary cold treatments for Li, Y., Wu, Y., Chen, H., Wu, J., & Li, Z. (2012). Population structure and oranges infested with Bactrocera invadens and Bactrocera zonata colonization of Bactrocera dorsalis (Diptera: Tephritidae) in China, in‐ (Diptera: Tephritidae) by comparison with existing cold treatment ferred from mtDNA COI sequences. Journal of Applied Entomology, schedules for Ceratitis capitata (Diptera: Tephritidae). Journal of 136(4), 241–251. https://doi.org/10.1111/j.1439-0418.2011.01636.x Economic Entomology, 106(4), 1608–1612. Li, Z. W., Li, X., Yu, Q. Y., Xiang, Z. H., Kishino, H., & Zhang, Z. (2009). The Hoffmann, A. A., & Ross, P. A. (2018). Rates and patterns of laboratory s m a l l h e a t s h o c k p r o t e i n ( sHSP) genes in the silkworm, Bombyx mori, and adaptation in (mostly) insects. Journal of Economic Entomology, 111(2), comparative analysis with other insect sHSP genes. BMC Evolutionary 501–509. https://doi.org/10.1093/jee/toy024 Biology, 9(1), 215. https://doi.org/10.1186/1471-2148-9-215 Hoffmann, A. A., Sørensen, J. G., & Loeschcke, V. (2003). Adaptation of Liu, H., Hou, B., Zhang, C., He, R., Liang, F., Gu, M., … Ma, J. (2014). Drosophila to temperature extremes: Bringing together quantitative Oviposition preference and offspring performance of the oriental and molecular approaches. Journal of Thermal Biology, 28(3), 175– fruit fly Bactrocera dorsalis and guava fruit fly B. correcta (Diptera: 216. https://doi.org/10.1016/S0306-4565(02)00057-8 Tephritidae) on six host fruits. Acta Ecologica Sinica, 9, 2274–2281. Hombach, A., Ommen, G., MacDonald, A., & Clos, J. (2014). A small heat Liu, H., Zhang, C., Hou, B. H., Ou‐Yang, G. C., & Ma, J. (2017). Interspecific shock protein is essential for thermotolerance and intracellular sur‐ competition between Ceratitis capitata and two Bactrocera spp. vival of Leishmania donovani. Journal of Cell Science, 127, 4762–4773. (Diptera: Tephritidae) evaluated via adult behavioral interference https://doi.org/10.1242/jcs.157297 under laboratory conditions. Journal of Economic Entomology, 110(3), Hu, J. T., Chen, B., & Li, Z. H. (2014). Thermal plasticity is related to 1145–1155. the hardening response of heat shock protein expression in two Liu, X., Jin, Y., & Ye, H. (2013). Recent spread and climatic ecological niche Bactrocera fruit flies. Journal of Insect Physiology, 67, 105–113. https:// of the invasive guava fruit fly, Bactrocera correcta, in mainland China. doi.org/10.1016/j.jinsphys.2014.06.009 Journal of Pest Science, 86(3), 449–458. https://doi.org/10.1007/ Huang, H. J., Xue, J., Zhuo, J. C., Cheng, R. L., Xu, H. J., & Zhang, C. s10340-013-0488-8 X. (2017). Comparative analysis of the transcriptional responses Liu, X., & Ye, H. (2009). Effect of temperature on development and sur‐ to low and high temperatures in three rice planthopper species. vival of Bactrocera correcta (Diptera: Tephritidae). Scientific Research Molecular Ecology, 26(10), 2726–2737. https://doi.org/10.1111/ and Essay, 4(5), 467–472. mec.14067 Lockwood, B. L., Julick, C. R., & Montooth, K. L. (2017). Maternal load‐ Huang, L. H., & Kang, L. (2007). Cloning and interspecific altered expres‐ ing of a small heat shock protein increases embryo thermal tolerance sion of heat shock protein genes in two leafminer species in response in Drosophila melanogaster. Journal of Experimental Biology, 220(23), to thermal stress. Insect Molecular Biology, 16(4), 491–500. https:// 4492–4501. doi.org/10.1111/j.1365-2583.2007.00744.x Lu, Y., Bai, Q., Zheng, X., & Lu, Z. (2017). Expression and enzyme ac‐ Huerta‐Cepas, J., Szklarczyk, D., Forslund, K., Cook, H., Heller, D., tivity of catalase in Chilo suppressalis (Lepidoptera: Crambidae) is re‐ Walter, M. C., … Kuhn, M. (2015). eggNOG 4.5: A hierarchical orthol‐ sponsive to environmental stresses. Journal of Economic Entomology, ogy framework with improved functional annotations for eukary‐ 110(4), 1803–1812. https://doi.org/10.1093/jee/tox117 otic, prokaryotic and viral sequences. Nucleic Acids Research, 44(D1), Lux, S. A., Copeland, R. S., White, I. M., Manrakhan, A., & Billah, M. K. D286–D293. (2003). A new invasive fruit fly species from the Bactrocera dorsa‐ Jang, E. B. (1991). Thermal death kinetics and heat tolerance in early and lis (Hendel) group detected in East Africa. International Journal of late third instars of the oriental fruit fly (Diptera: Tephritidae). Journal Tropical Insect Science, 23(4), 355–361. https://doi.org/10.1017/ of Economic Entomology, 84(4), 1298–1303. https://doi.org/10.1093/ S174275840001242X jee/84.4.1298 Malacrida, A., Gomulski, L., Bonizzoni, M., Bertin, S., Gasperi, G., & Kanehisa, M., Goto, S., Kawashima, S., Okuno, Y., & Hattori, M. (2004). Guglielmino, C. (2007). Globalization and fruitfly invasion and ex‐ The KEGG resource for deciphering the genome. Nucleic Acids pansion: The medfly paradigm. Genetica, 131(1), 1. https://doi. Research, 32(suppl_1), D277–D280. https://doi.org/10.1093/nar/ org/10.1007/s10709-006-9117-2 gkh063 Malewski, T., Bogdanowicz, W., Durska, E., Łoś, M., Kamiński, M., & Kawasaki, F., Koonce, N. L., Guo, L., Fatima, S., Qiu, C., Moon, M. T., … Kowalewska, K. (2015). Expression profiling of heat shock genes Ordway, R. W. (2016). Small heat shock proteins mediate cell‐au‐ in a scuttle fly Megaselia scalaris (Diptera, Phoridae). Journal of tonomous and‐nonautonomous protection in a Drosophila model Experimental Zoology Part A: Ecological Genetics and Physiology, for environmental‐stress‐induced degeneration. Disease Models & 323(10), 704–713. Mechanisms, 9(9), 953–964. Malmendal, A., Overgaard, J., Bundy, J. G., Sørensen, J. G., Nielsen, N. King, A. M., & MacRae, T. H. (2015). Insect heat shock proteins during C., Loeschcke, V., & Holmstrup, M. (2006). Metabolomic profiling of stress and diapause. Annual Review of Entomology, 60, 59–75. https:// heat stress: Hardening and recovery of homeostasis in Drosophila. doi.org/10.1146/annurev-ento-011613-162107 American Journal of Physiology‐Regulatory, Integrative and Comparative Klepsatel, P., Gáliková, M., Maio, N., Huber, C. D., Schlötterer, C., & Flatt, Physiology, 291(1), R205–R212. T. (2013). Variation in thermal performance and reaction norms Manjunatha, H., Rajesh, R., & Aparna, H. (2010). Silkworm thermal biol‐ among populations of Drosophila melanogaster. Evolution, 67(12), ogy: A review of heat shock response, heat shock proteins and heat 3573–3587. acclimation in the domesticated silkworm, Bombyx mori. Journal of Langmead, B., Trapnell, C., Pop, M., & Salzberg, S. L. (2009). Ultrafast and Insect Science, 10(1), 204. memory‐efficient alignment of short DNA sequences to the human Matsumura, T., Matsumoto, H., & Hayakawa, Y. (2017). Heat stress hard‐ genome. Genome Biology, 10(3), R25. https://doi.org/10.1186/ ening of oriental armyworms is induced by a transient elevation of gb-2009-10-3-r25 reactive oxygen species during sublethal stress. Archives of Insect Li, B., & Dewey, C. N. (2011). RSEM: Accurate transcript quantifica‐ Biochemistry and Physiology, 96(3). https://doi.org/10.1002/arch.21421 tion from RNA‐Seq data with or without a reference genome. BMC Myers, S. W., Cancio‐Martinez, E., Hallman, G. J., Fontenot, E. A., & Bioinformatics, 12(1), 323. https://doi.org/10.1186/1471-2105-12-323 Vreysen, M. J. (2016). Relative tolerance of six Bactrocera (Diptera: 1162 | GU et al.

Tephritidae) species to phytosanitary cold treatment. Journal of Sørensen, J. G., Nielsen, M. M., Kruhøffer, M., Justesen, J., & Loeschcke, Economic Entomology, 109(6), 2341–2347. V. (2005). Full genome gene expression analysis of the heat stress Nyamukondiwa, C., Kleynhans, E., & Terblanche, J. S. (2010). Phenotypic response in Drosophila melanogaster. Cell Stress & Chaperones, 10(4), plasticity of thermal tolerance contributes to the invasion potential 312–328. https://doi.org/10.1379/CSC-128R1.1 of Mediterranean fruit flies (Ceratitis capitata). Ecological Entomology, Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., & Kumar, S. 35(5), 565–575. https://doi.org/10.1111/j.1365-2311.2010.01215.x (2011). MEGA5: Molecular evolutionary genetics analysis using max‐ Nyamukondiwa, C., Terblanche, J. S., Marshall, K. E., & Sinclair, B. imum likelihood, evolutionary distance, and maximum parsimony J. (2011). Basal cold but not heat tolerance constrains plas‐ methods. Molecular Biology and Evolution, 28(10), 2731–2739. https:// ticity among Drosophila species (Diptera: Drosophilidae). doi.org/10.1093/molbev/msr121 Journal of Evolutionary Biology, 24(9), 1927–1938. https://doi. Tatusov, R. L., Galperin, M. Y., Natale, D. A., & Koonin, E. V. (2000). The org/10.1111/j.1420-9101.2011.02324.x COG database: A tool for genome‐scale analysis of protein func‐ Overgaard, J., Kristensen, T. N., Mitchell, K. A., & Hoffmann, A. A. tions and evolution. Nucleic Acids Research, 28(1), 33–36. https://doi. (2011). Thermal tolerance in widespread and tropical Drosophila org/10.1093/nar/28.1.33 species: Does phenotypic plasticity increase with latitude? American Vayssières, J. F., Carel, Y., Coubes, M., & Duyck, P. F. (2008). Development Naturalist, 178(S1), S80–S96. of immature stages and comparative demography of two cucur‐ Overgaard, J., Sørensen, J. G., Com, E., & Colinet, H. (2014). The rapid bit‐attacking fruit flies in Reunion Island: Bactrocera cucurbitae and cold hardening response of Drosophila melanogaster: Complex Dacus ciliatus (Diptera Tephritidae). Environmental Entomology, 37(2), regulation across different levels of biological organization. 307–314. Journal of Insect Physiology, 62, 46–53. https://doi.org/10.1016/j. Vázquez, D. P., Gianoli, E., Morris, W. F., & Bozinovic, F. (2017). Ecological jinsphys.2014.01.009 and evolutionary impacts of changing climatic variability. Biological Papadopoulos, N. T. (2008). Mediterranean fruit fly, Ceratitis capi‐ Reviews, 92(1), 22–42. https://doi.org/10.1111/brv.12216 tata (Wiedemann) (Diptera: Tephritidae). In J. L. Capinera (Ed.), Wang, H.‐J., Shi, Z.‐K., Shen, Q.‐D., Xu, C.‐D., Wang, B., Meng, Z.‐J., … Encyclopedia of Entomology (pp. 2318–2322). Dordrecht, The Wang, S. U. (2017). Molecular cloning and induced expression of six Netherlands: Springer. small heat shock proteins mediating cold‐hardiness in Harmonia axy‐ Permpoon, R., Aketarawong, N., & Thanaphum, S. (2011). Isolation and ridis (Coleoptera: Coccinellidae). Frontiers in Physiology, 8, 60. https:// characterization of Doublesex homologues in the Bactrocera species: doi.org/10.3389/fphys.2017.00060 B. dorsalis (Hendel) and B. correcta (Bezzi) and their putative pro‐ Wang, L., Feng, Z., Wang, X., Wang, X., & Zhang, X. (2009). DEGseq: An moter regulatory regions. Genetica, 139(1), 113–127. R package for identifying differentially expressed genes from RNA‐ Pieterse, W., Terblanche, J. S., & Addison, P. (2017). Do thermal toler‐ seq data. Bioinformatics, 26(1), 136–138. https://doi.org/10.1093/ ances and rapid thermal responses contribute to the invasion po‐ bioinformatics/btp612 tential of Bactrocera dorsalis (Diptera: Tephritidae)? Journal of Insect Weldon, C. W., Nyamukondiwa, C., Karsten, M., Chown, S. L., & Physiology, 98, 1–6. Terblanche, J. S. (2018). Geographic variation and plasticity in Qin, Y., Ni, W., Wu, J., Zhao, Z., Chen, H., & Li, Z. (2015). The potential climate stress resistance among southern African populations of geographic distribution of Bactrocera correcta (Diptera: Tephrididae) Ceratitis capitata (Wiedemann) (Diptera: Tephritidae). Scientific in China based on eclosion rate model. Applied Entomology and Reports, 8(1), 9849. https://doi.org/10.1038/s41598-018-28259-3 Zoology, 50(3), 371–381. Wellband, K. W., & Heath, D. D. (2017). Plasticity in gene transcription Qin, Y., Paini, D. R., Wang, C., Fang, Y., & Li, Z. (2015). Global establish‐ explains the differential performance of two invasive fish species. ment risk of economically important fruit fly species (Tephritidae). Evolutionary Applications, 10(6), 563–576. https://doi.org/10.1111/ PLoS ONE, 10(1), e0116424. eva.12463 Reitz, S. R., & Trumble, J. T. (2002). Competitive displacement among Willot, Q., Gueydan, C., & Aron, S. (2017). Proteome stability, insects and arachnids. Annual Review of Entomology, 47(1), 435–465. heat hardening and heat‐shock protein expression profiles in https://doi.org/10.1146/annurev.ento.47.091201.145227 Cataglyphis desert ants. Journal of Experimental Biology, 220(9), Ron, D., & Walter, P. (2007). Signal integration in the endoplasmic reticu‐ 1721–1728. lum unfolded protein response. Nature Reviews Molecular Cell Biology, Wos, G., & Willi, Y. (2018). Thermal acclimation in Arabidopsis lyrata: 8(7), 519. https://doi.org/10.1038/nrm2199 Genotypic costs and transcriptional changes. Journal of Evolutionary Schulze, S. K., Kanwar, R., Gölzenleuchter, M., Therneau, T. M., & Beutler, Biology, 31(1), 123–135. A. S. (2012). SERE: Single‐parameter quality control and sample Xie, C., Mao, X., Huang, J., Ding, Y., Wu, J., Dong, S., … Wei, L. (2011). comparison for RNA‐Seq. BMC Genomics, 13(1), 524. https://doi. KOBAS 2.0: A web server for annotation and identification of en‐ org/10.1186/1471-2164-13-524 riched pathways and diseases. Nucleic Acids Research, 39(Suppl 2), Shen, G. M., Huang, Y., Jiang, X. Z., Dou, W., & Wang, J. J. (2013). W316–W322. Effect of β‐cypermetherin exposure on the stability of nine house‐ Yang, Y., & Smith, S. A. (2013). Optimizing de novo assembly of short‐read keeping genes in Bactrocera dorsalis (Diptera: Tephritidae). Florida RNA‐seq data for phylogenomics. BMC Genomics, 14(1), 328. https:// Entomologist, 442–450. doi.org/10.1186/1471-2164-14-328 Sisodia, S., & Singh, B. N. (2006). Effect of exposure to short‐term heat Yuan, S., Kong, Q., Xiao, C., Yang, S., Sun, W., Zhang, J., & Li, Z. (2006). stress on survival and fecundity in Drosophila ananassae. Canadian Introduction to two kinds of artificial diets for mass rearing of Journal of Zoology, 84(6), 895–899. adult Bactrocera dorsalis (Hendel). Journal of Huazhong Agricultural Sørensen, J. G., Kristensen, T. N., & Loeschcke, V. (2003). The evolution‐ University, 25, 371–374. ary and ecological role of heat shock proteins. Ecology Letters, 6(11), Zizzari, Z. V., & Ellers, J. (2011). Effects of exposure to short‐term heat 1025–1037. https://doi.org/10.1046/j.1461-0248.2003.00528.x stress on male reproductive fitness in a soil arthropod. Journal Sørensen, J., & Loeschcke, V. (2001). Larval crowding in Drosophila of Insect Physiology, 57(3), 421–426. https://doi.org/10.1016/j. melanogaster induces Hsp70 expression, and leads to increased jinsphys.2011.01.002 adult longevity and adult thermal stress resistance. Journal of Żwirowski, S., Kłosowska, A., Obuchowski, I., Nillegoda, N. B., Insect Physiology, 47(11), 1301–1307. https://doi.org/10.1016/ Piróg, A., Ziętkiewicz, S., … Liberek, K. (2017). Hsp70 displaces S0022-1910(01)00119-6 small heat shock proteins from aggregates to initiate protein GU et al. | 1163

refolding. EMBO Journal, 36(6), 783–796. https://doi.org/10.15252/ embj.201593378 How to cite this article: Gu X, Zhao Y, Su Y, et al. A transcriptional and functional analysis of heat hardening in two invasive fruit fly species, Bactrocera dorsalis and Bactrocera SUPPORTING INFORMATION correcta. Evol Appl. 2019;12:1147–1163. https://doi. org/10.1111/eva.12793 Additional supporting information may be found online in the Supporting Information section at the end of the article.

Minerva Access is the Institutional Repository of The University of Melbourne

Author/s: Gu, X; Zhao, Y; Su, Y; Wu, J; Wang, Z; Hu, J; Liu, L; Zhao, Z; Hoffmann, AA; Chen, B; Li, Z

Title: A transcriptional and functional analysis of heat hardening in two invasive fruit fly species, Bactrocera dorsalis and Bactrocera correcta

Date: 2019-06-01

Citation: Gu, X., Zhao, Y., Su, Y., Wu, J., Wang, Z., Hu, J., Liu, L., Zhao, Z., Hoffmann, A. A., Chen, B. & Li, Z. (2019). A transcriptional and functional analysis of heat hardening in two invasive fruit fly species, Bactrocera dorsalis and Bactrocera correcta. EVOLUTIONARY APPLICATIONS, 12 (6), pp.1147-1163. https://doi.org/10.1111/eva.12793.

Persistent Link: http://hdl.handle.net/11343/249831

File Description: published version License: CC BY